ABSTRACT
Objective To establish the orthogonal partial least square （OPLS） model for the estimation of early postmortem interval （PMI） of asphyxial death rats in four ambient temperatures based on gas chromatography-mass spectrometry （GC-MS） metabolomics. Methods The 96 rats were divided into four temperature groups （5 ℃, 15 ℃, 25 ℃ and 35 ℃）. Each temperature group was further divided into 3 h, 6 h, 12 h and 24 h after death, and 6 other rats were taken as the control group. The cardiac blood was collected at the set time points for the four temperature groups and 0 h after death for the control group for the metabolomics analysis by GC-MS. By OPLS analysis, the variable importance in projection （VIP）>1 and the result of Kruskal-Wallis test P<0.001 were used to screen out the differential metabolite related to PMIs in the cardiac blood of rats of different temperature groups. Then OPLS regression models of different temperature groups were established with these metabolites. At the same time, a prediction group for investigating the prediction ability of these models was set up. Results Through the analysis of OPLS, 18, 15, 24 and 30 differential metabolites （including organic acids, amino acids, sugars and lipids） were screened out from the rats in groups of 5 ℃, 15 ℃, 25 ℃ and 35 ℃, respectively. The prediction results of the four temperature group models showed that the prediction deviation of 5 ℃ model was larger than that of other groups. The prediction results of other temperature groups were satisfactory. Conclusion There are some differences in the changes of metabolites in cardiac blood of rats at different ambient temperatures. The influence of ambient temperature should be investigated in the study of PMI estimation by metabolomics, which may improve the accuracy of PMI estimation.